Compile statement - warnings

[johncross]

I am currently trying to automate the loading of a df file overnight and have been able to do this succesfully. I also automatically compile programs overnight as well.

In certain cases I get a warning message with my compile command, does anyone know how I can prevent these messages from being shown and instead direct output to a log file

 

[Norman Biggar]

Something along the lines of

 [font=Courier New][size=2][font=Courier New][size=2]OUTPUT TO "compile.log".
[/size][/font]COMPILE [i]program[/i] SAVE NO-ERROR.
IF COMPILER:ERROR THEN DO:
[font=Courier New][size=2]  ASSIGN iCount = ERROR-STATUS:NUM-MESSAGES.
  DO iNum = 1 TO iCount:
	MESSAGE ERROR-STATUS:GET-MESSAGE(iNum).
  END.
END.
[/size][/font]
[/size][/font]

 

[toby.Harman]

Or the more high handed approach which is to not worry about what the error is and throw the program as hard as you can at the programmer who wrote it.

Beware - getting onto hobby horse and riding it!

The only "Warning" I am aware of that Progress supplies is Transaction Keyword given within actual transaction level. While I have lobbied hard within Progress to get this to be a compilation failure I am as yet unsuccessful. However this indicates that a transaction is not where the developer intended it to be by the use of the keyword TRANSACTION so there is a problem with the code!

 

[RobLyons]

The only "Warning" I am aware of that Progress supplies is Transaction Keyword given within actual transaction level

I regularly come across anothet type of Warning, after upgrading to V9.1D I have started seeing the warning "This code will not compile in earlier versions of progress" or something like that.
I have spent many sleepless nights worrying about this (well, OK I pondered on it for amout 1 and a half minutes while I was wating for the kettle to boil) it would be usefull if anyone knows how to supress it, but seeing as it is not exactl mission critical I haven't looked into it yet.

 

[toby.Harman]

How very useful of Progress to provide this!

It sounds like there is a keyword introduced in 9.1D which has been used and it is trying to warn you about the consequence of this.

compile no-error should fix this problem but you will need to check the compiler:error and compiler:stopped flags to confirm that the code did actually compile (as per Norman's response)

Or you can be nasty to your developers and put in a keyword forget so that they cannot use the specific feature! But that would be mean!